Which algorithm is commonly used for classification and can operate with margin-based optimization?

• A. Support vector machine
• B. K-means
• C. Naive Bayes
• D. Logistic regression

Answer: (A)

Margin-based optimization means choosing a decision boundary that maximizes the separation between classes. A support vector machine does exactly this: it finds a hyperplane that separates the classes with the largest possible margin, with the closest points on either side—the support vectors—defining that boundary. This focus on maximizing margin often leads to better generalization, and it can be extended to nonlinear boundaries using kernel tricks, giving both linear and nonlinear SVMs.

K-means is an unsupervised clustering method, not a classifier. Naive Bayes is a probabilistic classifier that relies on conditional independence assumptions, not margin optimization. Logistic regression estimates probabilities via maximum likelihood and does not optimize a geometric margin, even though it yields a linear decision boundary. So the algorithm that uses margin-based optimization for classification is the one based on support vector machines.